miRNAs and other post-transcriptional regulators are very well "genetic". They are encoded by genetic elements, are expressed and are affected by mutations. Just because this mode of regulation was not well known previously, it should not be classified as an epigenetic mechanism while the traditional protein based transcription factors (TF) are not.
Epigenetics, as it is originally defined (the "formal definition") is about mechanisms that can perpetuate the state of a cell to its next generation. Inheritance of gene expression programmes is therefore epigenetic. Although the gene expression programmes themselves can be implemented via different factors including protein and RNA based regulators, they would not necessarily constitute the epigenetic mechanisms that lead to inheritance of this state.
rg255's point of view is that any mechanism that causes a variation in the functional aspects of the genome without altering the genome sequence itself, would be epigenetic. This is technically correct but in that case all gene expression regulators including TFs should constitute epigenetic mechanisms.
Now, the main issue is where to draw the line between gene regulation and epigenetics?
In my opinion, the epigenetic mechanisms are one of the ways to regulate the gene expression. Although histone modifications and DNA methylation regulate gene expression and also confer heritability to the gene expression programme, the heritability can be implemented without them as well.
You can imagine a cell as a vessel which runs a system of biochemical reactions. This system can have multiple steady states (for e.g. multiple fates of a stem cell). To perpetuate a state, the new cell just needs to have the right initial conditions. This can be proved mathematically too. Such a system can be implemented via the traditional transcription factors as well. So what is epigenetic?
IMHO epigenetic was a loose term to denote something that people were not fully aware of, at that time. Anything that was not directly mediated by transcription factors was termed as epigenetic, including long distance regulators, non-coding RNA etc.
I would not classify non-coding RNAs as "epigenetic" for the very reason that they are encoded by genes and have more or less a direct effect on the target genes, just like TFs (which are apparently not epigenetic). As for the papers, there were many papers that used to assign these under epigenetic mechanisms, but that is IMO just too vaguely arbitrary. (Ironically, I happened to come across miRNAs and lncRNAs while I was doing a summer project on epigenetics and was reading relevant papers.)
What should be considered epigenetic would be a subject of another debate.